On Jan 18, 2006, at 21:57, Brian Grossman wrote:
On Wed, 18 Jan 2006 10:52:48 -0800
[EMAIL PROTECTED] (Justin Mason) wrote:
Andrew Pam writes:
Message-IDs MUST be globally unique as defined in the RFCs, so the
chance of collision is zero unless someone has very broken (or
malicious) email software.
A good 80% of email, according to recent statistics, is indeed broken
and/or malicious -- spam, malware, etc. I've seen many spams with
duplicate or missing Message-IDs.
Is anyone here blocking messages without message-ids? How's it
working for you?
Get many false positives?
As long as you don't do it for the clients that relay through your
server, it should be risk-free. All the big mail servers (sendmail,
qmail, exim, postfix, exchange) will add message IDs to all outgoing
messages. Watch out for malformed message IDs though, even the
support staff at MSN sends messages with weird message IDs.
Have you seen any spam that doesn't have message IDs?
-Johan
